Individual differences in drug dependence may be mediated by several abnormalities in associative learning, including perseveration of drug-seeking following contingency change, greater control over drug-seeking by Pavlovian stimuli, or greater sensitivity to drug reinforcement establishing higher rates of drug-seeking. To evaluate these three candidate markers for nicotine dependence, Experiment 1 contrasted daily (N = 22) and nondaily smoker groups (N = 22) on a novel instrumental learning task, where one S+ was first trained as a predictor of tobacco reward before being extinguished. Experiment 2 compared daily (N = 18) and nondaily smoker groups (N = 18) on a concurrent-choice task for tobacco and chocolate reward before an extinction test in which the tobacco response was extinguished, followed by a Pavlovian-to-instrumental transfer test, wherein the impact of tobacco and chocolate cues on concurrent choice was measured (gender was balanced within each smoker group). The results showed no group difference in sensitivity to extinction of either the stimulus-drug or response-drug contingency in Experiments 1 and 2, respectively, nor did groups show a difference in Pavlovian-to-instrumental transfer of control over tobacco choice. By contrast, nicotine-dependence status was marked by a higher frequency of tobacco choice in the concurrent-choice procedure, and this choice preference was associated with subjective craving (gender did not affect any behavioral measure). These results favor the view that nicotine dependence in this sample is not determined by individual predilection for perseveration or stimulus-control over drug-seeking, but by greater sensitivity to reinforcement of instrumental drug choice. Value-based decision theories of dependence are discussed.